Synthetically enhanced: unveiling synthetic data's potential in medical imaging research.

BACKGROUND: Chest X-rays (CXR) are essential for diagnosing a variety of conditions, but when used on new populations, model generalizability issues limit their efficacy. Generative AI, particularly denoising diffusion probabilistic models (DDPMs), offers a promising approach to generating synthetic images, enhancing dataset diversity. This study investigates the impact of synthetic data supplementation on the performance and generalizability of medical imaging research. METHODS: The study employed DDPMs to create synthetic CXRs conditioned on demographic and pathological characteristics from the CheXpert dataset. These synthetic images were used to supplement training datasets for pathology classifiers, with the aim of improving their performance. The evaluation involved three datasets (CheXpert, MIMIC-CXR, and Emory Chest X-ray) and various experiments, including supplementing real data with synthetic data, training with purely synthetic data, and mixing synthetic data with external datasets. Performance was assessed using the area under the receiver operating curve (AUROC). FINDINGS: Adding synthetic data to real datasets resulted in a notable increase in AUROC values (up to 0.02 in internal and external test sets with 1000% supplementation, p-value <0.01 in all instances). When classifiers were trained exclusively on synthetic data, they achieved performance levels comparable to those trained on real data with 200%-300% data supplementation. The combination of real and synthetic data from different sources demonstrated enhanced model generalizability, increasing model AUROC from 0.76 to 0.80 on the internal test set (p-value <0.01). INTERPRETATION: Synthetic data supplementation significantly improves the performance and generalizability of pathology classifiers in medical imaging. FUNDING: Dr. Gichoya is a 2022 Robert Wood Johnson Foundation Harold Amos Medical Faculty Development Program and declares support from RSNA Health Disparities grant (#EIHD2204), Lacuna Fund (#67), Gordon and Betty Moore Foundation, NIH (NIBIB) MIDRC grant under contracts 75N92020C00008 and 75N92020C00021, and NHLBI Award Number R01HL167811.

Secondary Intention Healing After Debridement for Bilateral Periorbital Necrotizing Fasciitis.

A 75-year-old male with a history of poorly controlled diabetes, hypertension, coronary artery disease, chronic obstructive pulmonary disease, and obesity presented with severe bilateral periorbital edema with necrosis and purulent discharge. Although hemodynamically stable, laboratory markers indicated systemic toxicity. Imaging showed bilateral periorbital edema extending into the frontal scalp, temporal fossa, and right orbit. Suspected to have necrotizing fasciitis, the patient underwent urgent debridement of bilateral upper and lower eyelids and was found to have postseptal extension of necrosis into the right orbit. During his hospitalization, he was treated with broad-spectrum antimicrobials and underwent a second surgery for exploration and debridement. The patient was lost to follow-up and found to have healed by secondary intention without any surgical reconstruction. Our case demonstrates not only a rare case of necrotizing fasciitis involving all 4 eyelids, but also an exceptional cosmetic and functional result after secondary intention healing.

The association of ASA score and outcomes following to.tal ankle arthroplasty.

BACKGROUND: This study seeks to evaluate the relationship between American Society of Anesthesiologist (ASA) score and postoperative outcomes following TAA. METHODS: The American College of Surgeons (ACS) National Surgical Quality Improvement Program (NSQIP) database was queried from 2007 to 2020 to identify 2210 TAA patients. Patients were stratified into low (n = 1328; healthy/mild systemic disease) or high (n = 881; severe/life-threatening systemic disease) ASA score cohorts. RESULTS: There was no statistically significant difference in complications, readmission, or reoperation rate based on ASA score. Increased ASA score was significantly associated with longer length of stay (low = 1.69 days, high = 1.98 days; p < .001) and higher rate of adverse discharge (low = 95.3 %, high = 87.4 %; p < .001). CONCLUSION: Higher ASA scores (3 and 4) were statically significantly associated with increased length of stay and non-home discharge disposition. These findings are valuable for physicians and patients to consider prior to TAA given the increased utilization of resources and cost associated with higher ASA scores. LEVEL OF EVIDENCE: Level III, Retrospective cohort study.

Surgical Approach and Body Mass Index Impact Risk of Wound Complications Following Total Hip Arthroplasty.

BACKGROUND: Previous studies have suggested that wound complications may differ by surgical approach after total hip arthroplasty (THA), with particular attention toward the direct anterior approach (DAA). However, there is a paucity of data documenting wound complication rates by surgical approach and the impact of concomitant patient factors, namely body mass index (BMI). This investigation sought to determine the rates of wound complications by surgical approach and identify BMI thresholds that portend differential risk. METHODS: This multicenter study retrospectively evaluated all primary THA patients from 2010 to 2023. Patients were classified by skin incision as having a laterally based approach (posterior or lateral approach) or DAA (longitudinal incision). We identified 17,111 patients who had 11,585 laterally based (68%) and 5,526 (32%) DAA THAs. The mean age was 65 years (range, 18 to 100), 8,945 patients (52%) were women, and the mean BMI was 30 (range, 14 to 79). Logistic regression and cut-point analyses were performed to identify an optimal BMI cutoff, overall and by approach, with respect to the risk of wound complications at 90 days. RESULTS: The 90-day risk of wound complications was higher in the DAA group versus the laterally based group, with an absolute risk of 3.6% versus 2.6% and a multivariable adjusted odds ratio of 1.5 (P < .001). Cut-point analyses demonstrated that the risk of wound complications increased steadily for both approaches, but most markedly above a BMI of 33. CONCLUSIONS: Wound complications were higher after longitudinal incision DAA THA compared to laterally based approaches, with a 1% higher absolute risk and an adjusted odds ratio of 1.5. Furthermore, BMI was an independent risk factor for wound complications regardless of surgical approach, with an optimal cut-point BMI of 33 for both approaches. These data can be used by surgeons to help consider the risks and benefits of approach selection. LEVEL OF EVIDENCE: Level III.

Most Publications Regarding Platelet-Rich Plasma Use in the Knee Are From Asia, Investigate Injection for Osteoarthritis, and Show Outcome Improvement: A Scoping Review.

PURPOSE: To evaluate and synthesize the available literature related to platelet-rich plasma (PRP) treatment of knee pathologies and to provide recommendations to inform future research in the field. METHODS: PubMed, CINAHL, and Scopus databases were queried on October 6, 2023. All identified citations were collated and uploaded into Covidence for screening and data extraction. Studies were included if they were human studies published in English with adult cohorts that received PRP as a procedural injection or surgical augmentation for knee pathologies with patient-reported outcome measures (PROMs) and level of evidence Levels I-IV. RESULTS: Our search yielded 2,615 studies, of which 155 studies from 2006 to 2023 met the inclusion criteria. Median follow-up was 9 months (±11.2 months). Most studies (75.5%) characterized the leukocyte content of PRP, although most studies (86%) did not use a comprehensive classification scheme. In addition, most studies were from Asia (50%) and Europe (32%) and were from a single center (96%). In terms of treatment, 74% of studies examined PRP as a procedural injection, whereas 26% examined PRP as an augmentation. Most studies (68%) examined treatment of knee osteoarthritis. Many studies (83%) documented significant improvements in PROMs, including 93% of Level III/IV evidence studies and 72% of Level I/II evidence studies, although most studies (70%) failed to include minimal clinically important difference values. The visual analog scale was the most-used PROM (58% of studies), whereas the Short Form Health Survey 36-item was the least-used PROM (5% of studies). CONCLUSIONS: Most published investigations of knee PRP are performed in Asia, investigate procedural injection for osteoarthritis, and show significant outcome improvements. In addition, this review highlights the need for better classification of PRP formulations. LEVEL OF EVIDENCE: Level IV, scoping Review of level I-IV studies.

Quantifying Imaging Agent Binding and Dissociation in 3-D Cancer Spheroid Tissue Culture Using Paired-Agent Principles.

Binding kinetics play an important role in cancer diagnosis and therapeutics. However, current methods of quantifying binding kinetics fail to consider the three-dimensional environment that drugs and imaging agents experience in biological tissue. In response, a methodology to assay agent binding and dissociation in 3-D tissue culture was developed using paired-agent molecular imaging principles. To test the methodology, the uptakes of ABY-029 (an IRDye 800CW-labeled epidermal growth factor receptor (EGFR)-targeted antibody mimetic) and IRDye-700DX carboxylate in 3-D spheroids were measured in four different human cancer cell lines throughout staining and rinsing. A compartment model (optimized for the application) was then fit to the kinetic curves of both imaging agents to estimate binding and dissociation rate constants of the EGFR-targeted ABY-029 agent. A statistically significant correlation was observed between apparent association rate constant (k3) and the receptor concentration experimentally and in simulations (r = 0.99, p < 0.05). A statistically significant difference was found between effective k3 (apparent rate constant of ABY-029 binding to EGFR) values for cell lines with varying levels of EGFR expression (p < 0.05), with no significant difference found between cell lines and controls for other fit parameters. Additionally, a similar binding affinity profile compared to a gold standard method was determined by this model. This low-cost methodology to quantify imaging agent or drug binding affinity in clinically relevant 3-D tumor spheroid models can be used to guide timing of imaging in molecular guided surgery and could have implications in drug development.

Evaluation of a Push-Pull Strategy for Spotted-Wing Drosophila Management in Highbush Blueberry.

We evaluated a novel push-pull control strategy for protecting highbush blueberry, Vaccinium corymbosum, against spotted-wing drosophila (SWD), Drosophila suzukii. Methyl benzoate (MB) was used as the pushing agent and a previously tested SWD attractive blend of lure-scents was used as the pulling agent. MB dispensers (push) were hung in the canopy and lure-scent dispensers (pull) were hung in yellow jacket traps filled with soapy water around the blueberry bushes. Blueberries were sampled weekly, and any infestation was inspected by examining the breathing tubes of SWD eggs which protrude through the skin of infested fruit. The frequency of infestation, i.e., the proportion of berries infested with at least one egg, and the extent of infestation, i.e., the mean number of eggs in infested berries, were significantly reduced in treatments receiving MB dispensers as a pushing agent when infestation rates were very high. However, the mass trapping devices as a pulling agent did not provide comparable protection on their own and did not produce additive protection when used in combination with the MB dispensers in push-pull trials. We conclude that MB has the potential to be implemented as a spatial repellent/oviposition deterrent to reduce SWD damage in blueberry under field conditions and does not require the SWD attractant as a pulling agent to achieve crop protection.

Improved intraoperative identification of close margins in oral squamous cell carcinoma resections using a dual aperture fluorescence ratio approach: first in-human results.

Significance: Surgical excision is the main treatment for solid tumors in oral squamous cell carcinomas, where wide local excision (achieving a healthy tissue margin of >5 mm around the excised tumor) is the goal as it results in reduced local recurrence rates and improved overall survival. Aim: No clinical methods are available to assess the complete surgical margin intraoperatively while the patient is still on the operating table; and while recent intraoperative back-bench fluorescence-guided surgery approaches have shown promise for detecting "positive" inadequate margins (<1 mm), they have had limited success in the detection of "close" inadequate margins (1 to 5 mm). Here, a dual aperture fluorescence ratio (dAFR) approach was evaluated as a means of improving detection of close margins. Approach: The approach was evaluated on surgical specimens from patients who were administered a tumor-specific fluorescent imaging agent (cetuximab-800CW) prior to surgery. The dAFR approach was compared directly against standard wide-field fluorescence imaging and pathology measurements of margin thickness in specimens from three patients and a total of 12 margin locations (1 positive, 5 close, and 6 clear margins). Results: The area under the receiver operating characteristic curve, representing the ability to detect close compared to clear margins (>5 mm) was found to be 1.0 and 0.57 for dAFR and sAF, respectively. Improvements in dAFR were found to be statistically significant (p<0.02). Conclusions: These results provide evidence that the dAFR approach potentially improves detection of close surgical margins.

Artificial intelligence-based three-dimensional templating for total joint arthroplasty planning: a scoping review.

PURPOSE: The purpose of this review is to evaluate the current status of research on the application of artificial intelligence (AI)-based three-dimensional (3D) templating in preoperative planning of total joint arthroplasty. METHODS: This scoping review followed the PRISMA, PRISMA-ScR guidelines, and five stage methodological framework for scoping reviews. Studies of patients undergoing primary or revision joint arthroplasty surgery that utilised AI-based 3D templating for surgical planning were included. Outcome measures included dataset and model development characteristics, AI performance metrics, and time performance. After AI-based 3D planning, the accuracy of component size and placement estimation and postoperative outcome data were collected. RESULTS: Nine studies satisfied inclusion criteria including a focus on computed tomography (CT) or magnetic resonance imaging (MRI)-based AI templating for use in hip or knee arthroplasty. AI-based 3D templating systems reduced surgical planning time and improved implant size/position and imaging feature estimation compared to conventional radiographic templating. Several components of data processing and model development and testing were insufficiently covered in the studies included in this scoping review. CONCLUSIONS: AI-based 3D templating systems have the potential to improve preoperative planning for joint arthroplasty surgery. This technology offers more accurate and personalized preoperative planning, which has potential to improve functional outcomes for patients. However, deficiencies in several key areas, including data handling, model development, and testing, can potentially hinder the reproducibility and reliability of the methods proposed. As such, further research is needed to definitively evaluate the efficacy and feasibility of these systems.

How are Oral Antibiotics Being Used in Total Joint Arthroplasty? A Review of the Literature.

While the role and benefit of perioperative intravenous (IV) antibiotics in patients undergoing total joint arthroplasty (TJA) is well-established, oral antibiotic use in TJA remains a controversial topic with wide variations in practice patterns. With this review, we aimed to better educate the orthopedic surgeon on when and how oral antibiotics may be used most effectively in TJA patients, and to identify gaps in the literature that could be clarified with targeted research. Extended oral antibiotic prophylaxis (EOAP) use in high-risk primary, aseptic revision, and exchange TJA for infection may be useful in decreasing periprosthetic joint infection (PJI) rates. When prescribing oral antibiotics either as EOAP or for draining wounds, patient factors, type of surgery, and type of infectious organisms should be considered in order to optimally prevent and treat PJI. It is important to maintain antibiotic stewardship by administering the proper duration, dose, and type of antibiotics and by consulting infectious disease when necessary.

Necroptosis does not drive disease pathogenesis in a mouse infective model of SARS-CoV-2 in vivo.

Necroptosis, a type of lytic cell death executed by the pseudokinase Mixed Lineage Kinase Domain-Like (MLKL) has been implicated in the detrimental inflammation caused by SARS-CoV-2 infection. We minimally and extensively passaged a single clinical SARS-CoV-2 isolate to create models of mild and severe disease in mice allowing us to dissect the role of necroptosis in SARS-CoV-2 disease pathogenesis. We infected wild-type and MLKL-deficient mice and found no significant differences in viral loads or lung pathology. In our model of severe COVID-19, MLKL-deficiency did not alter the host response, ameliorate weight loss, diminish systemic pro-inflammatory cytokines levels, or prevent lethality in aged animals. Our in vivo models indicate that necroptosis is dispensable in the pathogenesis of mild and severe COVID-19.

THA-Net: A Deep Learning Solution for Next-Generation Templating and Patient-specific Surgical Execution.

BACKGROUND: This study introduces THA-Net, a deep learning inpainting algorithm for simulating postoperative total hip arthroplasty (THA) radiographs from a single preoperative pelvis radiograph input, while being able to generate predictions either unconditionally (algorithm chooses implants) or conditionally (surgeon chooses implants). METHODS: The THA-Net is a deep learning algorithm which receives an input preoperative radiograph and subsequently replaces the target hip joint with THA implants to generate a synthetic yet realistic postoperative radiograph. We trained THA-Net on 356,305 pairs of radiographs from 14,357 patients from a single institution's total joint registry and evaluated the validity (quality of surgical execution) and realism (ability to differentiate real and synthetic radiographs) of its outputs against both human-based and software-based criteria. RESULTS: The surgical validity of synthetic postoperative radiographs was significantly higher than their real counterparts (mean difference: 0.8 to 1.1 points on 10-point Likert scale, P < .001), but they were not able to be differentiated in terms of realism in blinded expert review. Synthetic images showed excellent validity and realism when analyzed with already validated deep learning models. CONCLUSION: We developed a THA next-generation templating tool that can generate synthetic radiographs graded higher on ultimate surgical execution than real radiographs from training data. Further refinement of this tool may potentiate patient-specific surgical planning and enable technologies such as robotics, navigation, and augmented reality (an online demo of THA-Net is available at: https://demo.osail.ai/tha_net).

Whole Blood Metal Levels in the Setting of an Oncologic Endoprosthesis: Is There Cause for Concern?

BACKGROUND: Massive modular endoprostheses have become a primary means of reconstruction after oncologic resection of a lower extremity tumor. These implants are commonly made with cobalt-chromium alloys that can undergo wear and corrosion, releasing cobalt and chromium ions into the surrounding tissue and blood. However, there are few studies about the blood metal levels in these patients. QUESTION/PURPOSE: What is the whole blood cobalt and chromium ion level in patients with massive modular endoprostheses? METHODS: We performed a cross-sectional study of our total joints registry to identify patients with a history of an endoprosthetic reconstruction performed at our institution. Patients who were alive at the time of our review in addition to those undergoing an endoprosthetic reconstruction after an oncologic resection were included. Whole blood samples were obtained from 27 (14 male and 13 female) patients with a history of a lower extremity oncologic endoprosthesis. The median time from surgery to blood collection was 8 years (range 6 months to 32 years). Blood samples were collected and stored in metal-free ethylenediaminetetraacetic acid tubes. Samples were analyzed on an inductively coupled plasma mass spectrometer in an International Organization for Standardization seven-class clean room using polytetrafluoroethylene-coated instruments to reduce the risk of metal contamination. The analytical measuring range was 1 to 200 ng/mL for chromium and cobalt. Cobalt and chromium levels were considered elevated when the blood level was ≥ 1 ppb. RESULTS: Cobalt levels were elevated in 59% (16 of 27) of patients, and chromium levels were elevated in 26% (seven of 27). In patients with elevated metal ion values, 15 of 17 patients had a reconstruction using a Stryker/Howmedica Global Modular Replacement System implant. CONCLUSION: Blood metal levels were elevated in patients who received reconstructions using modular oncology endoprostheses Future work is needed to establish appropriate follow-up routines and determine whether and when systemic complications occur because of elevated metal levels and how to potentially address these elevated levels when complications occur. Prospective and retrospective collaboration between multiple centers and specialty societies will be necessary to address these unknown questions in this potentially vulnerable patient group. LEVEL OF EVIDENCE: Level IV, therapeutic study.

THA-AID: Deep Learning Tool for Total Hip Arthroplasty Automatic Implant Detection With Uncertainty and Outlier Quantification.

BACKGROUND: Revision total hip arthroplasty (THA) requires preoperatively identifying in situ implants, a time-consuming and sometimes unachievable task. Although deep learning (DL) tools have been attempted to automate this process, existing approaches are limited by classifying few femoral and zero acetabular components, only classify on anterior-posterior (AP) radiographs, and do not report prediction uncertainty or flag outlier data. METHODS: This study introduces Total Hip Arhtroplasty Automated Implant Detector (THA-AID), a DL tool trained on 241,419 radiographs that identifies common designs of 20 femoral and 8 acetabular components from AP, lateral, or oblique views and reports prediction uncertainty using conformal prediction and outlier detection using a custom framework. We evaluated THA-AID using internal, external, and out-of-domain test sets and compared its performance with human experts. RESULTS: THA-AID achieved internal test set accuracies of 98.9% for both femoral and acetabular components with no significant differences based on radiographic view. The femoral classifier also achieved 97.0% accuracy on the external test set. Adding conformal prediction increased true label prediction by 0.1% for acetabular and 0.7 to 0.9% for femoral components. More than 99% of out-of-domain and >89% of in-domain outlier data were correctly identified by THA-AID. CONCLUSIONS: The THA-AID is an automated tool for implant identification from radiographs with exceptional performance on internal and external test sets and no decrement in performance based on radiographic view. Importantly, this is the first study in orthopedics to our knowledge including uncertainty quantification and outlier detection of a DL model.

Anonymizing Radiographs Using an Object Detection Deep Learning Algorithm.

Radiographic markers contain protected health information that must be removed before public release. This work presents a deep learning algorithm that localizes radiographic markers and selectively removes them to enable de-identified data sharing. The authors annotated 2000 hip and pelvic radiographs to train an object detection computer vision model. Data were split into training, validation, and test sets at the patient level. Extracted markers were then characterized using an image processing algorithm, and potentially useful markers (eg, "L" and "R") without identifying information were retained. The model achieved an area under the precision-recall curve of 0.96 on the internal test set. The de-identification accuracy was 100% (400 of 400), with a de-identification false-positive rate of 1% (eight of 632) and a retention accuracy of 93% (359 of 386) for laterality markers. The algorithm was further validated on an external dataset of chest radiographs, achieving a de-identification accuracy of 96% (221 of 231). After fine-tuning the model on 20 images from the external dataset to investigate the potential for improvement, a 99.6% (230 of 231, P = .04) de-identification accuracy and decreased false-positive rate of 5% (26 of 512) were achieved. These results demonstrate the effectiveness of a two-pass approach in image de-identification. Keywords: Conventional Radiography, Skeletal-Axial, Thorax, Experimental Investigations, Supervised Learning, Transfer Learning, Convolutional Neural Network (CNN) Supplemental material is available for this article. © RSNA, 2023 See also the commentary by Chang and Li in this issue.

The Matthew Effect: Prevalence of Doctor and Physician Parents among Ophthalmology Applicants.

Objective This article determines the prevalence of physician parents among ophthalmology residency applications. Design Retrospective, single-center cohort study. Subjects All applicants to the University of Kentucky Ophthalmology Residency between 2018 and 2023. Methods Residency applications were reviewed, with data collection including applicant gender, self-identified Under-Represented in Medicine (URiM) status, United States Medical Licensing Examination (USMLE) Step 1 score, USMLE Step 2 score, and whether the application identified a doctor or physician as a parent. Doctor was defined as a profession requiring a doctorate degree, and similarly, physician as a profession requiring a medical degree. Results A total of 2,057 applications were reviewed, representing 54% of all match participants during the study period. Fourteen percent (296) of applications indicated a parent was a doctor and 12% (253) a parent was a physician. There were no differences between gender, URiM, USMLE Step 1, and Step 2 scores between applicants indicating a doctor or physician as a parent and those that did not ( p all > 0.4 and Cohen's d all < 0.02). Of the type of doctors, 85% (253) were physicians, 6% (17) optometrists, 6% (17) Doctors of Philosophy, 3% (8) dentists, 1% (1) pharmacist, and 1% (1) veterinarian. Eighty-six percent (217) of applications with a physician parent provided the type of physician, with ophthalmologist the most common (93, 43%). Ninety-eight percent (249) of applications with a physician parent provided the gender of the parent, with father (168, 68%) more common than mother (42, 17%) or both parents (39, 16%). Conclusion Physician parents are substantially overrepresented in ophthalmology residency applicants. This raises concerns regarding diversity and inclusion efforts for recruitment in medicine.

The Role of Ryanodine Receptor 2 in Drug-Associated Learning.

Type-2 ryanodine receptor (RyR2) ion channels facilitate the release of Ca 2+ from stores and serve an important function in neuroplasticity. The role for RyR2 in hippocampal-dependent learning and memory is well established and chronic hyperphosphorylation of RyR2 (RyR2P) is associated with pathological calcium leakage and cognitive disorders, including Alzheimer's disease. By comparison, little is known about the role of RyR2 in the ventral medial prefrontal cortex (vmPFC) circuitry important for working memory, decision making, and reward seeking. Here, we evaluated the basal expression and localization of RyR2 and RyR2P in the vmPFC. Next, we employed an operant model of sucrose, cocaine, or morphine self-administration (SA) followed by a (reward-free) recall test, to reengage vmPFC neurons and reactivate reward-seeking and re-evaluated the expression and localization of RyR2 and RyR2P in vmPFC. Under basal conditions, RyR2 was expressed in pyramidal cells but not regularly detected in PV/SST interneurons. On the contrary, RyR2P was rarely observed in PFC somata and was restricted to a different subcompartment of the same neuron - the apical dendrites of layer-5 pyramidal cells. Chronic SA of drug (cocaine or morphine) and nondrug (sucrose) rewards produced comparable increases in RyR2 protein expression. However, recalling either drug reward impaired the usual localization of RyR2P in dendrites and markedly increased its expression in somata immunoreactive for Fos, a marker of highly activated neurons. These effects could not be explained by chronic stress or drug withdrawal and instead appeared to require a recall experience associated with prior drug SA. In addition to showing the differential distribution of RyR2/RyR2P and affirming the general role of vmPFC in reward learning, this study provides information on the propensity of addictive drugs to redistribute RyR2P ion channels in a neuronal population engaged in drug-seeking. Hence, focusing on the early impact of addictive drugs on RyR2 function may serve as a promising approach to finding a treatment for substance use disorders.

A cadaveric breast cancer tissue phantom for phase-contrast X-ray imaging applications.

BACKGROUND: As mammography X-ray imaging technologies advance and provide elevated contrast in soft tissues, a need has developed for reliable imaging phantoms for use in system design and component calibration. In advanced imaging modalities such as refraction-based methods, it is critical that developed phantoms capture the biological details seen in clinical precancerous and cancerous cases while minimizing artifacts that may be caused due to phantom production. This work presents the fabrication of a breast tissue imaging phantom from cadaveric breast tissue suitable for use in both transmission and refraction-enhanced imaging systems. METHODS: Human cancer cell tumors were grown orthotopically in nude athymic mice and implanted into the fixed tissue while maintaining the native tumor/adipose tissue interface. RESULTS: The resulting human-murine tissue hybrid phantom was mounted on a clear acrylic housing for absorption and refraction X-ray imaging. Digital breast tomosynthesis was also performed. CONCLUSION: Both attenuation-based imaging and refraction-based imaging of the phantom are presented to confirm the suitability of this phantom's use in both imaging modalities.

Few-shot biomedical image segmentation using diffusion models: Beyond image generation.

BACKGROUND: Medical image analysis pipelines often involve segmentation, which requires a large amount of annotated training data, which is time-consuming and costly. To address this issue, we proposed leveraging generative models to achieve few-shot image segmentation. METHODS: We trained a denoising diffusion probabilistic model (DDPM) on 480,407 pelvis radiographs to generate 256 âœ• 256 px synthetic images. The DDPM was conditioned on demographic and radiologic characteristics and was rigorously validated by domain experts and objective image quality metrics (Frechet inception distance [FID] and inception score [IS]). For the next step, three landmarks (greater trochanter [GT], lesser trochanter [LT], and obturator foramen [OF]) were annotated on 45 real-patient radiographs; 25 for training and 20 for testing. To extract features, each image was passed through the pre-trained DDPM at three timesteps and for each pass, features from specific blocks were extracted. The features were concatenated with the real image to form an image with 4225 channels. The feature-set was broken into random patches, which were fed to a U-Net. Dice Similarity Coefficient (DSC) was used to compare the performance with a vanilla U-Net trained on radiographs. RESULTS: Expert accuracy was 57.5 % in determining real versus generated images, while the model reached an FID = 7.2 and IS = 210. The segmentation UNet trained on the 20 feature-sets achieved a DSC of 0.90, 0.84, and 0.61 for OF, GT, and LT segmentation, respectively, which was at least 0.30 points higher than the naively trained model. CONCLUSION: We demonstrated the applicability of DDPMs as feature extractors, facilitating medical image segmentation with few annotated samples.

A Stepwise Approach to Analyzing Musculoskeletal Imaging Data With Artificial Intelligence.

The digitization of medical records and expanding electronic health records has created an era of "Big Data" with an abundance of available information ranging from clinical notes to imaging studies. In the field of rheumatology, medical imaging is used to guide both diagnosis and treatment of a wide variety of rheumatic conditions. Although there is an abundance of data to analyze, traditional methods of image analysis are human resource intensive. Fortunately, the growth of artificial intelligence (AI) may be a solution to handle large datasets. In particular, computer vision is a field within AI that analyzes images and extracts information. Computer vision has impressive capabilities and can be applied to rheumatologic conditions, necessitating a need to understand how computer vision works. In this article, we provide an overview of AI in rheumatology and conclude with a five step process to plan and conduct research in the field of computer vision. The five steps include (1) project definition, (2) data handling, (3) model development, (4) performance evaluation, and (5) deployment into clinical care.